Limited effective gene flow between two interfertile red oak species

Abstract

Key message Highly elevated differentiation in different life stages between two interfertile oak species at a CONSTANS-like gene suggests a role of this gene in prezygotic isolation and adaptive divergence between species. Abstract Genome-wide differentiation patterns among oak species suggest that divergent selection can maintain species-specific adaptations and morphological integrity by reducing effective interspecific gene flow. While there is evidence for both pre-and post-zygotic isolation mechanisms in oaks (e.g., differences in flowering time, selection against hybrids), these mechanisms are rarely studied at each life stage from acorns to adult trees within the same forest. To assess the reproductive isolation mechanisms between two ecologically divergent species, we (1) quantified the number of hybrids in different life stages in Quercus rubra and Quercus ellipsoidalis, two interfertile red oaks with different adaptations to drought, and (2) assessed the timing of bud burst in both natural populations and in a seedling common garden trial. The low number of hybrids in all life stages suggested pre-zygotic isolation between species or selection in very early life stages that have not been sampled (e.g., early seed abortion). Significant differences in bud burst were found in two consecutive years between species in a common garden seedling trial but not in natural populations of the same provenance. In addition, we found evidence for divergent selection on several gene loci between species in each life stage. In particular, an SSR repeat located within the coding sequence of a CONSTANS-like gene, a locus involved in the photoperiodic regulation of flowering time and development, showed very high interspecific differentiation between species in all life stages (mean F ST = 0.83), compared to the average neutral differentiation of 3.7 %.

Adaptive Variation and Introgression of a CONSTANS-Like Gene in North American Red Oaks

“…Although slightly different marker sets were used in the genetic assignments, we have shown previously that very similar results are obtained using marker sets of 10, 16 or 44 markers [7,24]. The same method and criteria for genetic assignment analysis in the program STRUCTURE [37] were used as described in [25] and did not include the marker FIR013 or any other outlier marker.…”

“…Latitudinal and altitudinal gradients, as well as local environmental conditions such as water availability, have been shown to impact bud burst timing in other oak species [21,22]. Quercus rubra and Q. ellipsoidalis seedlings grown in a common garden exhibited differences in bud burst timing and leaf fall over two consecutive years [23] although this did not hold in natural populations of the same provenance [24]. Thus, differences in flowering time could still contribute to limit gene flow between these two species [24,25] and highly divergent markers, such as FIR013, located in the first exon of a COL gene, might be involved in their adaptive divergence and/or partial reproductive isolation.…”

“…Quercus rubra and Q. ellipsoidalis seedlings grown in a common garden exhibited differences in bud burst timing and leaf fall over two consecutive years [23] although this did not hold in natural populations of the same provenance [24]. Thus, differences in flowering time could still contribute to limit gene flow between these two species [24,25] and highly divergent markers, such as FIR013, located in the first exon of a COL gene, might be involved in their adaptive divergence and/or partial reproductive isolation. COL genes have been linked to adaptive divergence in other species, including the European white oak Q. petraea and Populus tremula L. [20,26,27].…”

Sharing of chloroplast haplotypes among red oak species suggests interspecific gene flow between neighboring populations

“…As in the European white oaks, these species revealed genomic divergence patterns as predicted by models of divergent selection with gene flow, with most markers showing low interspecific differentiation punctuated by gene loci with differentiation values (F ST values) largely above neutral expectations (outlier loci) Lind-Riehl et al 2014). For example, CONSTANS-like 1 was identified as outlier across multiple population pairs of Q. rubra and Q. ellipsoidalis, and was nearly fixed on alternative alleles in both species (F ST = 0.55-0.84), while the overall neutral interspecific differentiation was below 10% (F ST < 0.10) (Lind-Riehl et al 2014;Collins et al 2015). CONSTANS-like 1 is a candidate gene for flowering time (Yano et al 2000;Alberto et al 2013) and is involved in growth and development (Herrmann et al 2010;Hsu et al 2012).…”

“…Additionally, genetic assignment analysis at nuclear DNA markers (genomic and expressed sequence tagsimple sequence repeat (EST-SSR) markers) in the adult tree, seedling, and seed generation provided evidence for gene flow between Q. rubra and Q. ellipsoidalis, as shown by the presence of putative hybrids and introgressive forms that clustered with their parental species Collins et al 2015;Owusu et al 2015). Likewise, genetic assignment analyses at genomic and gene-based SSRs revealed evidence for considerable asymmetric gene flow between Q. ellipsoidalis and Q. velutina (Sullivan 2013;A.…”